Setting the group of a MR job

Imagine I am running a MapReduce job. The output in HDFS created is owned by that user – and is also in the user’s main group.
However if I want to to be in one of the other groups the user is in I don’t know how to do that.
Obviously I can chgrp the files afterwards, but that leaves a period of time when the files are in the wrong group. This could be important if I am using the groups for permissioning purposes.

Any ideas?

Thanks